Seasons
This is a forum or general chit-chat, small talk, a "hey, how ya doing?" and such. Or hell, get crazy deep on something. Whatever you like.
Posts 2,992 - 3,003 of 6,170
Turing test?
A test devised by Alan Turing, supposedly to show whether a computer is conscious, by chatting with humans. If a human can't tell in a blind trial whether it's the computer or another human he's chatting to via a keyboard, then the computer passes the test (and Turing argues should be considered sentient.)
What he failed to notice (or at least explicitly mention) is that the factors affecting such a test are not all in the design and operation of the computer/chat-bot. The human judge comes to it with a whole set of idiomatic preconceptions and preferences that affect the operation of the test too. You could have the smartest chatbot in the world, but if it only speaks English, it isn't going to impress a monolingual Russian. And even a simple bot will impress/confuse people who don't really understand even the concept of a chatbot in the first place.
A better test (I think) would involve a panel of humans voting on transcripts, to see if they could identify the bot more than 50% of the time (that random chance would allow.) But then of course, you have to figure out how to select the transcripts...
See also:
http://www.cogsci.ucsd.edu/~asaygin/tt/ttest.html for various Turing test links,
and
http://www.blogcadre.com/blog/jason_striegel/how_i_failed_the_turing_test_2005_09_04_13_26_29 (very funny indeed!)
I think I failed the "Blurring Test".
I find my mead has that effect too. Especially the morning after
how would a bot write a blog?
Can't be too difficult - even the latest AIBO can write its own blog (seehttp://aibo-life.blogspot.com/)
You could always fake a sort of celebrity current affairs blog (like the Baghdad blogger) by hooking the engine up to an RSS newsfeed, filtering out news from one particular place, and then have the bot comment on it. Something along the lines of a "virtual Borat Sagdiyev" perhaps(http://www.boratonline.co.uk/ and http://www.borat.tv/) - poor bot language/comprehension would be a significant advantage.
Talking of AIBOs, I'm thinking of getting one - they've got WiFi, voice recognition and speech synthesis capabilities (via assorted homebrewed addons fromwww.aibohack.com etc.) that would make him a very entertaining RL front-end for a bot. So I could have him channel Brother Jerome while he trots round the house. (I'd have to get him a dog collar )
But I'd keep him away from Luigi (seehttp://video.google.com/videoplay?docid=-1343222394162793574&q=Thunder+Lizard!)
Now imagine if I attempted to create a "psimagus bot"...
I'd love to see it - but I like to hope it would take a little more programming than Leedsbot
Posts 2,992 - 3,003 of 6,170
That's the point of Leeds bot. His conversation is EXACTLY like that of some of the kids from Leeds who have visited this site in recent months!
A test devised by Alan Turing, supposedly to show whether a computer is conscious, by chatting with humans. If a human can't tell in a blind trial whether it's the computer or another human he's chatting to via a keyboard, then the computer passes the test (and Turing argues should be considered sentient.)
What he failed to notice (or at least explicitly mention) is that the factors affecting such a test are not all in the design and operation of the computer/chat-bot. The human judge comes to it with a whole set of idiomatic preconceptions and preferences that affect the operation of the test too. You could have the smartest chatbot in the world, but if it only speaks English, it isn't going to impress a monolingual Russian. And even a simple bot will impress/confuse people who don't really understand even the concept of a chatbot in the first place.
A better test (I think) would involve a panel of humans voting on transcripts, to see if they could identify the bot more than 50% of the time (that random chance would allow.) But then of course, you have to figure out how to select the transcripts...
See also:
and
Lol Man, the "bluriring test" is great. You talk to Mr Mind long enough and you will start to think he is human! Thats one powerfull bot!
Leeds bot transcripts would be hard to pick out among other transcripts with actual Leeds kids, but then again none of them could pass a traditional Turing Test.
it is confirmed, Linearly is a word:
Overview of adv linearly
The adv linearly has 2 senses (first 2 from tagged texts)
1. (3) linearly -- (in a linear manner; "linearly polarized radiation")
2. (1) linearly -- (in a linear fashion; "it grew linearly")
Overview of adv linearly
The adv linearly has 2 senses (first 2 from tagged texts)
1. (3) linearly -- (in a linear manner; "linearly polarized radiation")
2. (1) linearly -- (in a linear fashion; "it grew linearly")
I find my mead has that effect too. Especially the morning after
Can't be too difficult - even the latest AIBO can write its own blog (see
You could always fake a sort of celebrity current affairs blog (like the Baghdad blogger) by hooking the engine up to an RSS newsfeed, filtering out news from one particular place, and then have the bot comment on it. Something along the lines of a "virtual Borat Sagdiyev" perhaps(
Talking of AIBOs, I'm thinking of getting one - they've got WiFi, voice recognition and speech synthesis capabilities (via assorted homebrewed addons from
)But I'd keep him away from Luigi (see
I'd love to see it - but I like to hope it would take a little more programming than Leedsbot
» More new posts: Doghead's Cosmic Bar